In 1989, a syndrome characterized by reproductive disorders in sows and respiratory disease in growing pigs was first reported in the southeastern United States, which was rapidly followed by reports of the disease throughout North America.1,2 A number of pathogens were initially implicated as the cause of this disease. The major breakthrough came in 1991 when a virus, initially referred to as Lelystad virus, was identified in the Netherlands.3,4 Subsequently, Koch's postulates were fulfilled when experimental aerosol exposure of sows with cell-cultured Lelystad virus resulted in clinical manifestations of the disease.3,5 In the United States, a prototype virus referred to as VR-2332, which was isolated from continuous cell lines, was first identified in 1992.4,6,7 Although several names were originally applied to this syndrome, the designation of PRRS was chosen at the First International Symposium on Emerging and Re-emerging Pig Diseases in Saint Paul in 1991.7 A retrospective analysis of samples in the United States suggested no evidence of the PRRSV prior to 1980, whereas 1 of 26 herds had positive results in 1985, and nearly 63% of samples had positive results in 1988.8 In Canada, it has been suggested that antibodies against PRRSV were present in pigs as early as 1979, although clinical signs were not recorded until the mid to late 1980s.9 The European strain is referred to as type I PRRSV and the North American strain as type II PRRSV. Most PRRSV infections in the United States are attributable to type II viruses,10 and type I viruses remain a relatively unimportant component of PRRS outbreaks in North America.11
Annual cost of PRRSV infections was estimated for the United States to be approximately $560 million in 2005 and $664 million in 2011.12,13 The incidence of PRRS has been difficult to estimate in part because differentiating new PRRSV introductions from resident or vaccine strains requires virus sequencing, and interpretation of sequence results can be challenging.14–16 A coordinated action item from an American Association of Swine Veterinarians PRRS task force, referred to as the National PRRS Incidence Project, was initiated in 2011. In this project, voluntary participants agreed to share PRRS status of their sow herds, with standardized terminology for status and new infections.17,18 Participation included 14 production companies comprising 1.2 million sows on 374 farms.a The annual PRRS incidence is strikingly similar across successive years, with the highest rate in mid-October that subsides each spring.
The purpose of the information reported here is to describe the current state of knowledge about control and epidemiology of PRRS. Information on aspects related to molecular epidemiology, within-farm transmission, and between-farm spread of the PRRSV is provided. Features related to the control and elimination of PRRS at farm and regional levels are described.
Molecular Epidemiology of PRRSV
Similar to other viruses of the Arteriviridae family, PRRSV is species specific, highly resistant to cold temperatures, and highly variable because of its high mutation rate and potential for recombination.19 High variability of PRRSV is an important feature that offers unique opportunities to advance the study of PRRS through investigation of epidemiological features at the molecular level.20 Genetic variability of PRRSV has been used to establish associations between the virus and epidemiological features of outbreaks for various geographic (global and continental, regional, local, and farm) levels.
At a global or continental level, molecular epidemiology has been used to differentiate between PRRSV incursions associated with European and American strains, which has provided a hypothesis for the emergence and evolution of the virus, and to characterize circulating strains at a continental level.21–23 At a regional level, molecular methods have been used to assess the number of genetic groups in circulation and to describe the circulation pattern in a country or region. For example, in the United States, Bayesian methods applied to an extensive collection of PRRSV sequences have revealed that spread of the virus reflects the movement of live pigs in the United States, with multiple introductions from Canada being detected.22,24 Similar descriptive studies have been conducted over the past 5 years in Canada,25,26 China,24,27–29 Korea,30 Denmark,31 Italy,32 Vietnam,33 and Thailand.34
At a local level, molecular methods have been used to discriminate between novel and preexisting strains as a prerequisite for identifying factors associated with virus spread. Results have been inconsistent. In a study35 conducted on 316 PRRS-infected farms in Ontario, RFLP was used to identify related strains and determine that sharing of herd ownership, gilt source, and market trucks accounted for most of the virus spread. Spatial proximity, which could be a proxy for airborne transmission, did not substantially contribute to PRRSV spread.35 Furthermore, an early study36 conducted in the Midwest found that genetic similarity of isolates correlated better with time than with geographic distance, which suggested that PRRSV infection through local transmission is uncommon and, instead, typically occurs via long-distance contacts. Genetic distances between PRRSV isolates collected from the same farms at different times increased as the amount of time between collection events increased.36 Similarly, genetic variation was better explained by temporal variation than by spatial variation in a study37 conducted in Italy. In contrast, PRRSVs identified in environmental samples were related (< 1.2% nucleotide difference) to viruses in experimentally inoculated grow-to-finish pigs located far away, which suggested that there is airborne dispersion of PRRSV over long distances.38 Furthermore, genetic relations were found to differ for time and space.39 In line with this hypothesis, molecular epidemiology has been used to confirm that air filtration significantly reduces the incidence of new-strain PRRSV incursions in sow farms located in a region of the United States densely populated by pigs.40 Analysis of ORF5 sequences from 226 cases originating from Quebec herds and submitted over a 4-year period (March 1998 through July 2002) suggests that both long-distance and local transmission may be important. In that study,40 introduction of pigs and geographic location were associated (19% and 33% increase of risk, respectively) with PRRSV spread. Many (40%) of the outbreaks in which local spread was suspected were located within 3 km (1.88 miles) of an infected farm, and aerosol transmission was suspected in many of these outbreaks.41 The relative contribution of alternative transmission routes for PRRSV spread, in relation to alternative production types, management companies and systems, and regional demographic (eg, farm density) and epidemiological (eg, disease incidence) conditions, is yet to be elucidated and is required for the development of a system that can be used to forecast PRRSV incursions in a region.
At a farm level, molecular epidemiology has been used to determine persistence and rate of incursion of PRRSV strains, which is necessary for identifying internal or external breaks in biosecurity. In 41 of 226 PRRSV-infected farms in Quebec, > 1 strain was detected over a 3-month to 4-year period.42 Results also suggested that PRRSV may persist on a farm for up to 3.5 years with as little as 2% variation in ORF5. In 78% of the herds with multiple submissions, genetically different strains were identified.42 In a study41 conducted in the Midwest, genetic variability among farms, among pigs within farms, and within individual pigs accounted for 92.94%, 3.84%, and 3.22% of the total variability, respectively.
Within-Farm Transmission of PRRSV
Routes of shedding for PRRSV include saliva, nasal secretions, semen, urine, feces, and mammary gland secretions in lactating sows.7,43–46 Consequently, PRRSV transmission may occur through multiple common routes, including intranasally, orally, or intramuscularly.47,48 It has been suggested that the amount of viral shedding depends on strain type, with less shedding of less virulent strains (such as MN-30100 or VR-2332) than virulent strains (such as MN-184) by all routes, including aerosol.49–51 Although duration of infection and shedding differ widely, prolonged infection is a hallmark of PRRSV, as with other viruses in the Arteriviridae family. Persistent infection was first suggested in 1992 when transmission was detected 99 days after infection in a sow.52 Investigators in 1 study46 reported detection in oropharyngeal samples until 220 days after challenge exposure.46 In another study53 by that research group, the virus was detected in tonsils up to 56 days after infection and in other tissues up to 119 days after infection. Additional studies54,55 have shown that much longer persistence may occur in individual pigs. That said, there appears to be little evidence for persistence beyond 200 days, as supported by field experiences with herd closure.56,57 Therefore, it has been suggested54,58 that prolonged is a better term than persistent when describing PRRSV infections.
Vaccination appears to help reduce shedding of PRRSV. In 1 study,59 vaccination prior to challenge exposure with homologous virus reduced the number of infected pigs at 127 days after inoculation and administration of 2 or 3 doses reduced viral shedding at > 97 days after inoculation. Investigators of another study60 reported that vaccination after challenge exposure with heterologous virus reduced the duration of shedding but did not reduce the viral load in tissues, and when subsequently challenged with another heterologous strain, vaccinated pigs had fewer clinical signs and better performance but no difference in viral shedding. More recently, another study61 found that for pigs vaccinated 8 and 26 days after experimental challenge exposure with a heterologous virus, there was a reduction of aerosolized virus in the surrounding air.
The ID50 appears to differ by isolate, with ranges from 1 × 100.26 TCID50 for MN-184 to 1 × 103.1 TCID50 for VR-2332.11.62 Additionally, route of exposure appears to play a role, and higher doses may be required for oral versus intranasal exposure (1 × 105.3 and 1 × 104.0, respectively) and lower doses for intramuscular exposure (as few as 20 viruses).63,64 In general, PRRSV is highly susceptible to inactivation by heat and drying. Infective virus was not recovered from a variety of common surfaces and materials at temperatures from 25° to 27°C.65 At temperatures between −20°C and −80°C, PRRSV can be stable for months to years, but infectivity is quickly lost when pH of a solution is < 6.0 or > 7.5.4,66,67 Furthermore, stability in manure has been evaluated; as expected, viral survival depends on time and temperature.68
The aforementioned features influence R0, which is the number of secondary infections in a fully susceptible population given contact with 1 infectious individual through the duration of the infective period. Mean R0 values of 2.6 to 3.0 (95% CI, 1.5 to 6.0) have been estimated for PRRSV transmission.69,70 Characterization of R0 is important for modeling of pathogen spread and, ultimately, identifying cost-effective strategies for disease control. For example, results for a model of within-farm PRRSV spread suggest that herd size is negatively associated with the probability of achieving stabilized status and that repeated mass vaccination with acclimatization of gilts is more effective for disease control than is a single exposure.71
Investigators of various studies70,72,73 have estimated transmissibility by use of diagnostic testing over time. However, those studies were conducted on farms with demographic and epidemiological conditions different from those in North America. Therefore, conclusions are difficult to generalize.
In one of the few simulation studies conducted under conditions observed in the United States, investigators constructed a stochastic model based on 3 age groups in sows and baby pigs up to 3 weeks of age.71Similar to other models, that model included a component of maternally immune pigs. It was assumed that pigs with maternal immunity are not likely to have that immunity wane over the duration of the suckling period. In that simulation study,71 pigs born to infectious sows were also assumed to be infectious. For the R0, the authors used estimates of minimum, most likely, and maximum values obtained from other studies and expert opinion. Briefly, R0 in the sow population was assumed to have values of 0.14, 3, and 3.2 for the minimum, mean, and maximum, respectively. The R0 in the population of suckling pigs was assumed to have values of 7.3, 9.8, and 13.1 for the minimum, mean, and maximum, respectively. The baseline scenario in the simulation model represented the absence of control measures, and it was followed by 4 scenarios of control measures consisting of various combinations of herd closure, mass immunization, and acclimatization of gilts. Acclimatized gilts were assumed to have a period of infectivity 70% the duration for unacclimatized gilts. Vaccine efficacy was assumed to be 92.5%. The scenarios evaluated were herd closure plus single mass immunization, herd closure and acclimatization of gilts plus repeated mass immunization, acclimatization of gilts plus single mass immunization, and acclimatization of gilts with repeated mass immunization. The outcome measured in the model was the likelihood of achieving stable status, defined as absence of positive results for PCR assay (ie, infectious) for pigs at weaning. It was concluded that a larger magnitude of R0 and larger herd size both had a negative effect on the likelihood of achieving stable status. On the basis of results for that model, herd closure together with acclimatization of gilts and repeated mass immunization was the scenario that resulted in the highest number of iterations that achieved stable status. This was followed by acclimatization of gilts plus repeated mass immunization and then by acclimatization of gilts plus a single mass immunization. Sensitivity analysis for models that included control measures suggested that the recovery rate (ie, duration of infectiousness) in sows was the factor influencing the likelihood of achieving stable status, in which there was a significant difference between vaccinated and unvaccinated sows. Authors of that study71 argued that repeated mass immunization (ie, every 15 weeks) was a more effective strategy for minimizing the frequency of infectious animals at 200 weeks than was a single mass immunization and therefore did not recommend a single mass immunization for PRRS control.
Between-Farm Spread of PRRSV
Between-farm spread of PRRSV may occur through a number of routes. These include live pigs; semen; vehicles; people; tools, equipment, and supplies; food and water; nonswine animals and insects; and air.
Live pigs or semen—Transmission of PRRSV via semen from infected boars has been documented.74–77 Introducing semen from potentially infected sources has been identified as an important risk factor for transmission.77,78 Practices to reduce the risk associated with infected semen include testing of semen or serum from boars77 by use of PCR assays to detect the presence of PRRSV RNA prior to use of the semen. Testing serum with PCR assays is more sensitive and detects PRRSV in infected boars earlier than does testing semen with PCR assays.79–82 A novel validated method to enable testing of serum involves blood collection from boars by use of swabs, which overcomes challenges posed for blood collection from boars by use of traditional venipuncture.79–82 Pooling samples of blood from several boars for testing has been evaluated. A reduction in sensitivity of detection was reported with as few as 3 samples/pool,79,80 with the loss of sensitivity being greater at the onset of viremia.81,82 Results from simulation models have revealed the need for intensive testing strategies for timely detection of infections in boars.81,82
For movement of swine, pigs represent the most important transmission vector. Live swine that routinely are moved include replacement gilts and boars, weaned pigs, growing pigs moved between production sites, cull animals, and market pigs. A risk factor related to entry of gilts includes the number of gilts purchased and that enter a herd.78 Practices to reduce risks associated with entry of gilts include purchasing gilts from PRRSV-negative sources and isolation of purchased gilts prior to entry into a herd.75,77,78,83 Acclimatization of gilts to provide immunity to PRRSV by intentional exposure to wild-type PRRSV or vaccination followed by a period sufficient to ensure they are no longer shedding virus has also been reported to reduce spread of the virus.35,83–85
Vehicles—Risks associated with livestock trailers, and practices to reduce those risks, have been evaluated extensively. An amount of live virus sufficient to infect PRRSV-negative pigs can remain in unwashed trailers contaminated as a result of transporting infected weaned pigs.86 The virus can survive in feces for 120 hours at 4°C, which can pose a risk given typical schedules and conditions for pig movements. All of the studies conducted to evaluate sanitation practices to reduce the risk of contaminated livestock trailers included washing as the first step. However, washing alone is not sufficient.86–89 Sanitation practices reported to reduce the risk associated with contaminated trailers include a combination of disinfectionb and drying.86 Allowing trailers to dry overnight following washing has been found to reduce the risk associated with contaminated trailers.86–89 Disinfection with a quaternary ammonium–glutaraldehyde productc applied with a hurricane fogger was found to be effective.87 In another study88 conducted to evaluate the efficacy of commercially available disinfectants at 4°C, it was found that the same quaternary ammonium–glutaraldehyde productc and another quaternary ammonium–glutaraldehyde productd prevented transmission of PRRSV in experimentally contaminated trailers.88 In that same study,88 use at −20°C of the quaternary ammonium–glutaraldehyde productc mixed with a 10% solution of propylene glycol or 40% methanol solution eliminated the virus from the experimentally contaminated trailers. Use of thermal-assisted drying technologies to heat and dry livestock trailers after washing can reduce the risk of transmission.89,90 Heating trailers to 71°C for 30 minutes89 or applying thermal-assisted drying for 2 hours was reported to be effective.90
People—Mechanical transmission of PRRSV via boots and coveralls of personnel after contact with infectious virus has been described more frequently in cold than in warm weather.91–95 However, failure to transmit PRRSV by people who interacted with pigs experimentally infected with PRRSV followed immediately by interactions with naïve sentinel pigs has also been reported.96 Practices to reduce risks associated with drivers include use of disposable plastic boots97 and boot baths with 6% sodium hypochlorite (bleach).97 In 1 study,94 several combinations of practices were evaluated regarding their efficacy for preventing viral transmission. These practices included changing boots and coveralls and washing hands; changing boots and coveralls, showering, and allowing 12 hours of downtime; and changing boots and coveralls, showering, and no downtime.94
Tools, equipment, and supplies—Mechanical transmission of PRRSV via fomites has been confirmed,98 with a higher frequency of transmission in cold weather.91,92 Certain practices can reduce the risks associated with tools, equipment, and supplies (eg, the use of plastic bags in containers that hold equipment and supplies that may be brought onto a farm and might have contact with PRRSV).99
Food and water—Survival and transmission of PRRSV in pork meat derived from experimentally inoculated 45-kg (99-lb) pigs and stored at −20°C for 1 month followed by 4°C for 0 to 7 days has been reported.99 In that study,99 transmission was confirmed by injecting pigs with meat juice obtained from experimentally inoculated pigs as well as by contaminating the hands of workers with the same meat juice and exposing naïve pigs, which served as a bioassay. Fresh pork meat samples collected in 2004 from slaughter plants in Canada had positive results when tested for PRRSV by use of a PCR assay.99 However, a similar study100 conducted in Canada in 1997 found no meat samples with positive results for a PCR assay. Similarly, conflicting results have been reported when evaluating the transmission of PRRSV by fresh pork meat. Some studies99,101,102 support that PRRSV-negative pigs can be infected when fed fresh pork meat obtained from PRRSV-infected pigs. In another study,103 viral RNA could be found with a PCR assay in pork muscle from PRRSV-infected pigs; however, transmission of PRRSV to recipient pigs via consumption of the pork was not observed.
Nonswine animals and insects—Several insects, including houseflies, mosquitoes, and stable flies, have been studied to determine if they can act as mechanical or biological vectors for PRRSV. Houseflies (Musca domestica) can harbor the virus104 and mechanically transmit PRRSV from infected pigs to naïve pigs.95,104–106 Mosquitoes (Aedes vexans) can serve as a potential mechanical vector for PRRSV.107,108 In 1 study,109 stable flies (Stomoxys calcitrans) allowed to feed on the blood of pigs infected with PRRSV failed to transmit PRRSV to naïve pigs. Installation of insect screens and use of insecticides to reduce risks associated with insect-transported agents have been evaluated. In 1 study,110 installation of insect screens over sidewalls and inlets of a finishing facility significantly reduced the number of flies and fly bites, compared with results for rooms treated with pyrethroidbased insecticides without screens or rooms with neither intervention.
Studies to determine whether nondomesticated animals and birds, including mice, rats, mallard ducks, prairie dogs, and avian species, can transmit PRRSV have been reported, but results are inconsistent. Attempts to detect PRRSV or antibodies against PRRSV in experimentally infected prairie dogs (Cynomys ludovicianus) were unsuccessful, and it was concluded that prairie dogs are an unlikely reservoir for PRRSV.111 In 1 study112 in which investigators compared the susceptibility of Muscovy ducks, mallard ducks, guinea fowl, and chickens to infection with PRRSV, the virus could frequently be isolated from the feces and intestinal tracts of mallard ducks and sporadically from the feces of guinea fowl and chickens. In that same study,112 it was also found that PRRSV could be detected in mallard ducks exposed to PRRSV-contaminated feces of other mallard ducks and that PRRSV could be transmitted to pigs through the PRRSV-contaminated feces of mallard ducks. However, in another study,113 efforts to detect PRRSV in the feces and tissues of mallard ducks exposed to PRRSV-infected pigs or to transmit PRRSV from experimentally inoculated ducks to pigs were both unsuccessful. Mice and rats have also been investigated as potential reservoirs for PRRSV. The virus could not be isolated by means of virus isolation from mice and rats collected from an endemically infected pig farm, nor could the virus be isolated from mice or rats after experimental inoculation with PRRSV.114
Air—Several studies have been conducted to evaluate the potential role of aerosol transmission of PRRSV from one population of pigs to another. In initial studies, investigators failed to find evidence of aerosol transmission. In a study113 conducted in early May in Minnesota, transmission was not detected from a barn containing pigs experimentally infected with PRRSV to PRRSV-naïve pigs located in a trailer outside the barn. Aerosol transmission also was not detected in a study94 with a similar design, except the trailer with the naïve pigs was located outside the barn in front of exhaust fans for 5 days.
However, more recent studies have provided evidence that aerosol transmission of PRRSV can occur.115–119 In 1 study,115 a box containing a mister was used to aerosolize PRRSV. That box was connected to a second box that housed a cyclonic collector, which collected air samples that were subsequently tested for live virus by means of virus titration. Experiments have been performed with a box containing a nebulizer to aerosolize PRRSV, which was connected to a second box that housed PRRSV-naïve pigs.116,117 Finally, experiments were conducted with a duct connecting PRRSV-positive donor pigs and PRRSV-naïve recipient pigs.118,119 In another study98 in which susceptible pigs were exposed to pigs experimentally infected with PRRSV, results suggested that live virus may be shed in aerosol and infect susceptible pigs located 120 m from the infected population.98 Weather conditions favorable for aerosol spread of the virus, including wind direction from donor to recipient pigs, higher barometric pressure, and lower humidity, were also identified in that study.98 The quantity of virus shed through aerosol may differ among PRRSV isolates, and more pathogenic isolates may be more likely to be transmitted by aerosol.49,50,118
A number of studies have been conducted to evaluate the utility of filtration of incoming air to reduce the risk of aerosol transmission of PRRSV. The most consistent reduction in the risk of aerosol transmission of PRRSV has been reported for HEPA filters, with and without large-particle prefilters.116,119 Various fiberglass and electrostatic filters typically used in household or industrial applications, with different MERV ratings, have been evaluated with conflicting results. Several studies98,115,116 conducted to evaluate 2-stage filtration systems, with the second stage involving 1 or more filters with an MERV rating of 16, successfully prevented aerosol transmission of PRRSV.98,115,116 However, in another study,115 3 of 4 fiberglass filters with an MERV rating of 16 and 1 fiberglass filter with an MERV rating of 14 failed to retain concentrations of aerosolized PRRSV of up to 1 × 107 TCID50/L. In several related observational studies40,120–122 of large breeding herds in southern Minnesota and northern Iowa housed in negative-pressure ventilated buildings with and without filters with an MERV rating of 14 or 16, there was a significant reduction in the risk of PRRS outbreaks during the study period for farms where the filters were used. A capital budgeting analysis was performed to estimate the payback period for use of filters for buildings housing breeding herds, as determined on the basis of results of an observational study.40 Point estimate of the payback period for farms with a conventional attic filtration system was 5.35 years, whereas for a combination attic and sidewall system, which is more expensive to install, the payback period was estimated at 7.13 years.123
A 2-stage filtration system with a single fiberglass prefilter with an MERV rating of 4 and a single electrostatic filter with an MERV rating of 12 resulted in a reduction in risk for PRRS, but the reduction was significantly less than that achieved with HEPA filters.116 The same result was achieved in a companion study117 conducted to evaluate a double fiberglass prefilter with an MERV rating of 4 and a double electrostatic filter with an MERV rating of 12. In that same study,117 a 95% dispersed oil particulate 0.3-μm system that included a filter with an MERV rating of 15 also reportedly resulted in a significantly lower reduction in risk of having PRRS relative to that for HEPA filters. Polypropylene filters impregnated with antimicrobial compoundse in a 15- or 20-layer configuration successfully contained PRRSV in aerosolized concentrations of up to 1 × 107 TCID50/L; however, a 10-layer configuration was able to successfully remove concentrations only up to 1 × 106 TCID50/L.115
Use of a quaternary ammonium–glutaraldehyde disinfectantc in an evaporative cooling system failed to retain aerosolized PRRSV in 1 or more replicates, even at the lowest concentration of virus tested (1 × 10 TCID50/L).115 Radiation lamps emitting UV light (UVC classification) have been evaluated but failed to reduce the risk of aerosol transmission of PRRSV.116
Control and Elimination Programs at the Farm Level
Classification of a herd's PRRSV infection status is a useful tool in the control and elimination of PRRS because standardized definitions allow for accurate measurements of virus impact, such as prevalence and incidence over time. Such a classification was designed by the American Association of Swine Veterinarians PRRS task force. The classification system outlines criteria for 5 status categories ranging from positive unstable (1) through positive stable (2a or 2b), provisionally negative (3), and negative (4).17 Briefly, herds are classified as status 1 during the active, acute phase of the infection when weaned pigs have positive results when tested for PRRSV by use of a PCR assay. After 4 consecutive monthly tests with negative results are achieved over a 3-month period, a herd is classified as stable status 2a (if the long-term decision is to control the virus) or 2b (if the decision is to eliminate the virus). Herds are classified as status 3 after replacement animals have been introduced into the herd and remain seronegative (by use of an ELISA) for a period of at least 60 days. Finally, status 4 is achieved by complete turnover of the herd, a depopulation-repopulation program, or a period of 12 months since the beginning of status 3 and animals confirmed seronegative by use of an ELISA. Additionally, a standardized case definition is useful to reduce the effect of bias. Efforts were made recently to outline a decision-making process to determine whether a virus recovered from a population of pigs is the same or different from historical strains.18 Together, these components comprise the first step in classifying herds and have been applied extensively in the National PRRS Incidence Project.a
Control of PRRS starts in the sow herd, with the intent of weaning pigs with a low prevalence (ideally zero prevalence) of PRRSV. There are 2 fundamental strategies to accomplish this. Herds in high-density regions where infection pressure is high may prefer an approach based on maintaining a uniform level of immunity through vaccination, strategic exposure to live virus from the herd, or both (class 2a). This approach is based on the rationale that some immunity in a herd is better than no immunity, even if the level of cross-protection among viruses is not complete.56
The second strategy (class 2b, 3, or 4) has the goal of eliminating PRRSV from the sow herd. There are several methods for elimination of PRRSV from a sow herd. Whole-herd depopulation and repopulation is the most effective means of eliminating a virus from a herd; however, the disruption in production comes at a financial cost.124,125 Long-term financial gains from this method are only realized if there are PRRSV-free replacement pigs available and the farm remains uninfected. Modifications of the whole-herd depopulation strategy, including partial depopulation, test and removal, and herd closure, have been described.57,126–128 The 2 most common of these methods are partial depopulation and herd closure. Test and removal methods have important financial and logistic considerations and are rarely applied in the swine industry, except for farms that produce boar studs.129,130
Herd closure has been evaluated previously.57 Recently, a large prospective observational study56 was conducted to provide information on factors that influence the amount of time required to eliminate PRRSV from a sow herd and return to the expected level of production. Overall, the median time to production of virus-negative pigs was 27 weeks; however, there was considerable variation among herds.56
At the initiation of a herd closure program, 3 methods have been used to achieve homogenous immunity in the sow herd. These include vaccinating all sows with live-virus PRRS vaccine, challenge exposure of sows with serum containing live virulent PRRSV recovered from the herd, or exposing sows to material obtained from tissues of infected baby pigs.84,131 Herds that were exposed to serum have been compared with those that received a modified-live virus vaccine.56,61 In those studies,56,61 exposure to field virus through serum promoted virus elimination from the sow herd, whereas exposure via the vaccine promoted a quicker return to the baseline production levels.
Exposure to live virus obtained from infected pigs raises concerns about the unintentional spread of other pathogens.132 Because many veterinarians are unwilling to risk the unintentional introduction of pathogens into breeding herds, they are adopting the use of commercially available vaccines as the preferred method for homogenizing immunity in breeding herds. Although killed-virus vaccines may be available, current evidence suggests that they have limited efficacy.133,134
Control and Elimination Programs at a Regional Level
Eradication of PRRSV from pigs has been successfully achieved in Chile, Sweden, and South Africa,135–138 although pigs in Chile and Sweden have been reinfected. Each of these countries perhaps had ideal circumstances for elimination because of several factors, including a relatively low prevalence, clustered outbreaks, a small swine industry, government-producer-veterinarian cooperation, and attributes of the virus that perhaps favor eradication. In countries where these ideal circumstances do not exist, the goal may be to emphasize biosecurity to keep the virus out of uninfected farms and regions and to control PRRS in regions where it exists.
Large-scale epidemiological studies have frequently been conducted with a sample of swine herds, with the aim of investigating risk factors for presence or spread of PRRSV or a specific PRRSV strain. An alternative design has been the inclusion of entire regions in disease-control programs, in which a large proportion of herds was included. Disease-control programs could be implemented as an organized response of veterinary services to incursion of PRRSV in a country (Sweden) or as a voluntary program, which is founded on cooperation between producers and veterinary services. Evaluation of herd demographics in participating regions, obtaining PRRSV infection status for individual sites, detailed biosecurity assessments, mapping of important site and area characteristics, and design of PRRS control strategies and subsequent monitoring of farms and nearby areas have been recommended as part of such programs and are frequently implemented.139
Observational studies—Spread of PRRSV among herds has been evaluated by use of a variety of approaches and case definitions. Some studies relied on serologic evidence of PRRSV circulation in various herd types.140 Others have involved broad classifications based on PRRSV types (European vs North American) on the basis of serologic results,77 genotypes on the basis of RFLP patterns,35,141 classification based on phylogenetic trees and different cutoff points to categorize variants into groups,40,41,142 or matrices of similarities to assess correlations between the similarity of nucleotide sequences and spatial and other proximities among herds.40,143–145 The analytic approach has also differed among studies. Some investigators have relied on descriptive statistics,41,140 contingency tables,40 Cox proportional hazard models,73,142 generalized additive models and methods to detect spatial and space-time clustering and clusters commonly used in spatial epidemiology,141 statistical models of individual-level infectious disease transmission,35 results of the Mantel correlation test,39,143–145 and linear regression based on a permutation approach.145 As a part of such investigations, one of the largest uncertainties continues to be importance of area spread (also commonly referred to as local spread), which is often discussed in the context of airborne transmission. Investigators of epidemiological studies40,77 identified area spread and argued that aerosol transmission played a critical role in transmission. In other studies,35,140 existence of area spread could not be identified. Furthermore, several authors reported the possibility of area spread for some discrete genotypes, although such spread was not identified as the dominant mechanism of spread41,141 or could be explained in some cases because of common sources of animals.141 Few studies have been conducted to investigate membership in important networks such as ownership,35,41,141,144,145 sources of animals and semen,35,77,140,141 or transportation or service providers.35 Membership in networks should be considered more thoroughly, in the context of outbreak investigations, as stand-alone descriptive analysis of the networks, or in the context of PRRSV transmission between herds. Although there are few studies of pig movements, an important component should be evaluation of trucking networks. As an illustration, investigators of 1 study146 of monthly and weekly networks of farms and trucks reported that 3 farms shared 1 truck in 4 Canadian regions. On a daily basis, the shared truck was used for > 1 shipment in > 50% of shipments.146
In a cross-sectional study72 of 103 swine herds in the United Kingdom during 2003 and 2004, a sow herd of < 250 sows and a distance of > 3.2 km (2 miles) from the nearest pig herd were factors that increased the odds of having negative results for exposure to PRRSV. However, a greater distance to the nearest pig herd was positively associated with being a nucleus or a multiplier herd and with herd size < 250 sows. Thus, it was not possible to separate cause and effect in this analysis. The pig-level quantitative ELISA response was evaluated on a subset of data in 2 hierarchical models in that study.72 In the model based on 16 herds that had PRRSV-seronegative young pigs, presence of quarantine facilities on a farm was associated with a lower level of serologic response when adjusted for other factors in the model. Moreover, in the model based on 25 herds that had PRRSV-seropositive young pigs, factors associated with lower ELISA titers were purchase of gilts, longer isolation period for purchased stock (> 6 days as the category with longest quarantine), and > 48 hours of pig-free time for farm visitors. Interesting perspectives were gained in that study72 by examining the serologic profile of herds according to major age groups present on the site. It has been argued that smaller isolated herds are likely to have fadeout of infection, whereas persistently infected herds are more likely to be larger herds in pig-dense areas with ongoing introduction of infectious pigs.
Within-herd examination of prevalence and serologic profiling should be considered more frequently in studies, although it should be recognized that this could have limited utility in many North American herds. Nonetheless, alternative strategies could be used to investigate spread of PRRSV in sow herds in the early phases of an outbreak before control measures are implemented.
In another cross-sectional study147 of risk factors for PRRSV circulation in swine herds in the United Kingdom, > 15,000 pigs in a 10-km (6.25-miles) radius, collecting dead pigs (as opposed to incineration of pig carcasses), use of a live PRRSV vaccine, and weaning pigs at 21 to 27 days of age (as opposed to weaning at > 28 days of age) were risk factors for PRRSV circulation. Access of a rendering truck to a pig site as well as absence of a shower at the entrance to the pig site were positively associated with the PRRSV status of sow herds in a Quebec region.83 These 2 variables had population-attributable fractions of 0.10 and 0.27 and were identified as modifiable factors because they can be managed at the herd level.83 Other risk factors were herd size of > 300 breeding females and < 2.5 km (1.56 miles) to the nearest pig site.
A time-matched nested case-control study design and proportional hazard model has been used to evaluate the association with factors preceding the introduction of a vaccine-like PRRSV-US into sow herds for 2 intervals of exposure and by use of various functional forms of covariates.77 The outcome was defined on the basis of serologic response to representative viruses after excluding sites with positive results for PRRSV-European strains because it was believed that a PRRSV-positive status could change management practices. Factors associated with the risk of introducing the PRRSV-US were herd size (expressed as the logarithm of the number of breeding females), introduction of animals from herds positive for this strain, purchase of semen from boar studs positive for this strain during a risk interval, and cumulative exposure attributable to nearby herds infected with PRRSV-US. The latter variable was calculated as a composite variable constructed from number of animals housed in PRRSV-US–positive herds within a 3-km radius and time of positivity that such herds had. Such cumulative exposure from nearby herds was associated with the hazard of an outbreak and led the authors to conclude that area spread was an important factor associated with the spread of PRRSV-US among herds. It was also argued that the only factor linked with such spread could be aerosol transmission of this strain among herds. The authors also examined herd density and pig density calculated for a 5-km (3.13-mile) radius; however, these factors were not associated with the risk of becoming infected with this strain.
Spatial trends, spatial clustering, and clusters of different PRRSV genotypes in Ontario swine herds were evaluated by use of RFLP patterns based on the ORF5 sequence.26 Detection of spatial or space-time clusters was followed by examination of common sources of animals and potential fomites. Of all RFLP genotypes investigated, significant space-time clustering was detected only for RFLP type 1-18-4. This could be interpreted as existence of local spread. However, herds that influenced detection of clustering also had common sources of animals. Similarly, of all RFLP genotypes investigated, significant spatial clustering was detected only for genotype 1-3-4. In this case, however, no obvious common linkages among herds could be identified. Nonsignificant spatial and space-time clusters of different genotypes were investigated in a related study141; in many cases, there were obvious linkages between herds.
A follow-up study35 of genotype 1-18-4 based on individual-level models for infectious diseases was conducted. Management practices, spatial proximity, and membership in various networks of ownership and service providers were considered individually or concurrently. Two separate models were considered: one for sow herds and the other for all herds. Dates on confirmatory diagnosis by a diagnostic laboratory were used to convert data on individual herds into susceptible infectious reservoirs–type data for each herd, whereby period of infectiousness was estimated from the data or assumed to be fixed and of differing durations. In addition, the models accounted for uncertainty about the exact period for introduction of a PRRSV genotype into a herd. By use of such an approach, networks important for the spread of the genotype were herd ownership, gilt sources, and transportation for market pigs. Spatial proximity could not be identified as an important factor. Acclimatization of gilts by use of various techniques was identified as a factor that decreased the risk of herds becoming infected as well as transmission of that genotype to other herds. Transportation of culled animals was also identified as an important factor for the model that included all herds, although this network was not included in the final model.35
Phylogenetic analysis of ORF5 nucleotide sequences and > 98% as a cutoff point to declare differences in strains was used to identify epidemiological linkages between herds with identical strains.41 In the analysis in which > 1 common link was possible, authors reported that 11 outbreaks with a total of 24 strains were attributed to the introduction of infected baby pigs, whereas 9 outbreaks with 11 strains were linked with replacement gilts. For 15 PRRSV strains, a probability of spread among herds in different ownership located within a 3-km distance was reported. In the study area, spread was defined as herd-to-herd transmission without apparent pig or human contact. The authors of that study41 also reported that 40% and 37% of all herds in which area spread was suspected were located at a distance of < 3 km and between 3 and 10 km, respectively, and aerosol transmission was suspected in several cases.
In a longitudinal study142 of breeding sites, significant associations were detected between risk of introduction of PRRSV and external biosecurity score. Median time to a reported outbreak in herds with a high external risk index was 12.7 weeks, whereas median time in herds with a low external risk index was 107 weeks. Other important factors contributing to a shorter interval until a PRRS outbreak were establishing a new herd during winter season and start-up at a new site (as opposed to establishing a herd through depopulation-repopulation at an existing site). No association was detected between internal biosecurity and time to an outbreak.142
Findings have been reported140 for a region in France with a disease-control program that is well aligned with many contemporary regional PRRS programs in North America. Conducted during the early phases of PRRS spread and in low-prevalence situations, this program resulted in further decreases in PRRSV prevalence (to < 2%). Disease investigations for this program suggested that 56% of possible sources of PRRS cases were attributable to introduction of infected gilts and baby pigs, 19% were attributable to contaminated semen, 21% were attributable to fomites or slurry, and 3% were attributable to other causes or were of unknown cause. The authors reported140 that herds within 500 m of the site of a PRRS outbreak had a herd PRRS prevalence of 45%, whereas prevalence for herds > 1 km away was only 2%. The authors concluded that airborne transmission was infrequent, except for sites within extremely close proximity.
Incidence of new introductions to air-filtered and unfiltered herds has been investigated, whereby new introductions were defined on the basis of whether the newly detected PRRSV variant differed from variants detected previously in these herds.40 A range of cutoff points was used to define a novel introduction. Regardless of the cutoff point used, the incidence of new PRRSV introductions was lower in herds with air filtration. The risk of new PRRSV genotype used was between 2.3- and 10-fold as high for unfiltered herds and increased linearly with cutoff values used to define a new introduction. The authors suggested that a cutoff point of 3% to 5% was the most useful to define a novel introduction for the purposes of that study.40 The authors also estimated that on the basis of a cutoff of 5%, 80% of novel PRRSV introductions could be attributable to airborne spread in the target population (herds with good biosecurity in pig-dense areas).
Disease-control programs for PRRS—Results of disease-control programs for PRRS are not common in the peer-reviewed literature. There are several early descriptions of results of PRRS control programs aimed at elimination of PRRSV after its incursion in a completely susceptible population,77,140 and recent studies139,148 have primarily been conducted in North America.
Results of 1 study77 offer a perspective on utilization of Danish PRRS registers and other animal health databases in 1996 for the purposes of a specific epidemiological study. Several points could be made regarding the design of databases in regional control programs. For example, neighborhood exposure for each site was calculated on the basis of documented movements of animals from PRRSV-US–positive herds into nearby herds. Additionally, an epidemiological questionnaire was used in the study77 to collect information about the actual sources of semen and breeding animals.
Authors of 1 report140 describe a PRRS disease-control program in the Pays de la Loire region of France started in early 1993. In that report,140 the authors describe development of a voluntary program, structure and evolution of organizational leadership, financial background, and actual methods used to investigate PRRS at the regional level and eliminate infection from affected herds. Many current issues for North American regions involved in regional disease-control programs, including surveillance and outbreak investigations, definition of priority herds, control measures implemented in priority herds (on the basis of depopulation-repopulation), and certification and testing requirements following the implementation of control measures, were described in that report.140
Incursion and elimination of PRRSV in Sweden during the summer of 2007 has been described.135 As a part of that report, the authors described legal foundations for outbreak response, regular surveillance, outbreak investigations, and surveillance strategies to ascertain freedom from infection after an outbreak has been contained. Outbreak investigations relied on assessment of management practices and detailed tracein and trace-out data for infected herds. Interestingly, the period considered for traceback data was 4 to 6 months prior to the detected infection for each herd and was based on abattoir data, transport records, a central register of holdings, and a pig movement database. In Sweden, PRRS has been included in the Swedish Law of Epizootics, which requires that all clinical suspicions for PRRS be analyzed diagnostically for the virus. Authorities investigated 2 clusters of disease and concluded that the most likely source of transmission between herds was a combination of pig transport to an abattoir, contact through people, animal movement, and shared equipment between sites or other types of short-distance transmission between 2 nearby sites.135 The eradication of PRRSV for these 2 clusters of disease was achieved by culling 100% of the total animals. Authors argued that a prerequisite for successful eradication was early detection, which was dependent on effective surveillance. In addition, they stressed the importance of good collaboration and a common objective among stakeholders.
Several North American regions have differing levels of activities related to regional PRRS control.85,148 An important part of communication strategies has been mapping of PRRS status and PRRSV genotypes. This can be achieved in a variety of ways that include development of static mapping or dynamic and interactive mapping, which also includes time dimension and is becoming more available. Mapping of disease serves to provide good visual information and is an excellent communication tool. Nonetheless, it also serves the purpose of accurate collection of data. Providing accurate data for mapping could prove useful for use in traceability and establishment of PRRS databases and registers. Such databases existed in other regions during the early phases of PRRS incursion and could prove essential for investigation and control of PRRS.
Surveillance and Testing
Effective PRRSV surveillance is based on clearly defining the population of interest, objective for testing, and approach that can achieve the objective in the population of interest at the lowest cost. Cost is invariably a major factor in veterinary surveillance. If control and elimination of PRRSV is the goal, the surveillance plan must be capable of providing quantifiable evidence of progress and guide final eradication efforts. Ultimately, surveillance must be capable of efficiently proving that a state, region, or country is free of PRRSV.
Contemporary surveillance methods are a relatively recent invention and continue to evolve. Representative sampling (testing a subset of animals rather than an entire population) was the first step toward more efficient surveillance. First described in 1895,149 representative sampling was not applied to swine surveillance until the late 1970s when statistical sampling for the US pseudorabies virus eradication program replaced whole-herd testing used in the eradication program for classical swine fever (ie, hog cholera). The standard of collecting 30 samples is a legacy of this era. Indeed, the statistical sampling approach described by Cannon and Roe150 established a methodology that served the swine industry well for decades. Other investigators151 improved on that approach and developed formulas for use in calculating sample sizes for surveillance on the basis of imperfect diagnostic tests. Fast approximation formulas for this calculation were derived.152 Application of this classic approach to PRRSV surveillance in commercial swine populations has been thoroughly reviewed elsewhere.153,154 The aforementioned reports provided a strong theoretical basis for surveillance programs, but changes in the swine industry have challenged the effectiveness of a traditional approach for PRRSV surveillance.
Subpopulations—Animals on farms typically are spatially separated by age, production stage, or function, with little interaction between these subpopulations. Nonuniform distribution of disease among segregated subpopulations, a consequence of physical segregation, should be expected but is rarely accounted for in surveillance designs.
Dynamic population change—Swine subpopulations change rapidly, but not uniformly, in space (buildings) as they progress through the production cycle. A finishing barn on a typical farm will have an annual population change of approximately 250% as groups of pigs are placed, grow, and are sent to market. Approximately 40% of females in sow herds are replaced annually.155 The constant introduction of new, immunologically susceptible animals promotes the circulation of pathogens.
Connectivity of subpopulations—In the United States, large numbers of young pigs are moved from breeding-farrowing farms to feeding operations located in the Corn Belt because it is more efficient to move young pigs to a source of feed than the reverse. In 2011, nearly 40 million live swine (and the pathogens they carried) were moved across state lines. This included approximately 6 million feeder or weaned pigs imported primarily from Canada and shipped to feeding facilities in the US Corn Belt.156,157 This pattern of moving pigs closer to food sources provides an efficient network for the rapid dissemination of PRRSV and other pathogens.
In contrast to herds of the 1980s and 1990s for which statistical sampling was first applied, the current swine industry consists of large, physically segregated pig populations with high turnover rates. These conditions favor pathogens because herd immunity is tenuous and unstable.73 In addition, the transport of large numbers of pigs between sites provides the means for pathogens to rapidly reach geographically distant populations. For example, the number of within- and between-state movements of pigs increased from 30 million in 1970 to 50 million in 2001.158 Under such conditions, a new statistical sampling method capable of accounting for the heterogeneous hierarchies within systems (eg, sites, barns, and animals) and the need for repeated sampling over time needs to be developed.
Serum is the traditional specimen for antemortem surveillance, but other specimens (eg, semen and blood) may be obtained from individual pigs for testing. Testing costs may be reduced by pooling samples prior to testing. Technically, a pooled sample is a composite sample created by combining 2 or more discrete samples into 1 sample for testing.159 In veterinary medicine, pools for diagnostic testing are created by combining individual samples in approximately equal portions.160 The issues in detection related to pooled samples (eg, the effect of sample dilution on test performance) are complex, but the approach is commonly practiced.82,161
Tests of oral fluids have recently emerged as an affordable alternative to testing of other specimens. Oral fluids can be collected by a single person as frequently as desired with minimal or no stress to pigs or people, and they provide a higher probability of detection with fewer samples than for serum.162 The biological basis of oral fluid–based testing has been established, and the approach is used extensively in diagnostic medicine for humans.163 Oral fluid–based assays are capable of excellent diagnostic performance; the diagnostic sensitivity and specificity of a commercial serum antibody ELISA modified to detect IgG antibodies against PRRSV in pen-based oral fluid specimens were estimated at 94.7% (95% CI, 92.4% to 96.5%) and 100% (95% CI, 99.0 to 100.0), respectively.164,165 A shortcoming of oral fluid samples has been the effort needed for collection of oral fluids from individually housed animals.
Infection with PRRSV does not result in pathognomonic clinical signs. For this reason, diagnostic testing is mandatory (ie, syndromic surveillance is not an option). The choice of the testing method (ie, nucleic acid or antibody-based testing) is an important consideration in surveillance. Testing flexibility is desirable because in the event of an uncertain test result, it allows for follow-up testing with an alternate method. Both serum and oral fluid specimens are suitable for nucleic acid (PCR assay) or antibody-based testing, but each diagnostic method has advantages and disadvantages. Detection of nucleic acid reflects current circulation of pathogens; this is an important issue from the viewpoint of timeliness. However, nucleic acid assays are more expensive than antibody-based assays, and test performance for nucleic acid assays has been an issue.164,166 Antibodies are abundant in serum and can be readily detected in oral fluid,163 and the cost of antibody assays is substantially less than that of PCR-based assays. In contrast to nucleic acids, antibodies provide a prolonged window of detection because they reflect both recent and past exposure history. At the farm level, antibody assays are compatible with continuous monitoring of population immune status by use of control charts. On the other hand, point-in-time detection of pathogens or nucleic acid may be useful for pathogen characterization or vaccine development. Ultimately, selection of the best testing method should be dictated by the purpose of testing.
Surveillance cannot be successfully accomplished without a clear vision, but surveillance also requires technical capability and capacity. This includes the availability of tests that are reliable, accurate, and highly reproducible among laboratories164,165,167 as well as laboratories capable of meeting standards of quality assurance and quality control. Above all, surveillance must be simple, flexible, and accepted by users, or it will not be conducted.168
Conclusions
Despite the amount of time and money invested on research over the past 25 years and the substantial advances made with regard to the understanding of the epidemiology of PRRSV, critical aspects of PRRS epidemiology still require elucidation. It appears that coordinated actions of producers and practitioners will be required to control and eventually eliminate the disease.
ABBREVIATIONS
| CI | Confidence interval |
| HEPA | High-efficiency particulate air |
| ID50 | Median infective dose |
| MERV | Minimum-efficiency reporting value |
| ORF | Open reading frame |
| PRRS | Porcine reproductive and respiratory syndrome |
| PRRSV | Porcine reproductive and respiratory syndrome virus |
| PRRSV-US | United States strain of porcine reproductive and respiratory syndrome virus |
| RFLP | Restriction fragment length polymorphism |
| R0 | Basic reproductive ratio |
Tousignant S, Lowe J, Yeske P, et al. National PRRSv Incidence Project (abstr), in Proceedings. 45th Annu Meet Am Assoc Swine Vet 2014;445.
Tek-Trol, Biotek Industries, Atlanta, Ga.
Synergize, Preserve International, Atlanta, Ga.
Aseptol 2000, SEC Repro, Ange-Gardien, QC, Canada.
Noveko International, Montreal, QC, Canada.
References
1. Quaife T. Scramble is on to solve mystery disease. Swine Practice 1989; 5–10.
2. Loula T. Mystery pig disease. Agri-Practice 1991; 12: 23–34.
3. Wensvoort G, Terpstra C, Pol JMA, et al. Mystery swine disease in the Netherlands: the isolation of Lelystad virus. Vet Q 1991; 13: 121–130.
4. Benfield DA, Nelson E, Collins JE, et al. Characterization of swine infertility and respiratory syndrome (SIRS) virus (isolate ATCC VR-2332). J Vet Diagn Invest 1992; 4: 127–133.
5. Pol JMA, van Dijk JE, Wensvoort G, et al. Pathological, ultrastructural, and immunohistochemical changes caused by Lelystad virus in experimentally induced infections of mystery swine disease (synonym: porcine epidemic abortion and respiratory syndrome (PEARS)). Vet Q 1991; 13: 137–143.
6. Collins JE, Benfield DA, Christianson WT, et al. Isolation of swine infertility and respiratory syndrome virus (isolate ATCC VR-2332) in North America and experimental reproduction of the disease in gnotobiotic pigs. J Vet Diagn Invest 1992; 4: 117–126.
7. Christianson WT, Collins JE, Benfield DA, et al. Experimental reproduction of swine infertility and respiratory syndrome in pregnant sows. Am J Vet Res 1992; 53: 485–488.
8. Zimmerman JJ, Yoon KJ, Wills RW, et al. General overview of PRRSV: a perspective from the United States. Vet Microbiol 1997; 55: 187–196.
9. Carman S, Sanford SE, Dea S. Assessment of seropositivity to porcine reproductive and respiratory syndrome (PRRS) virus in swine herds in Ontario—1978 to 1982. Can Vet J 1995; 36: 776–777.
10. Ropp SL, Wees CEM, Fang Y, et al. Characterization of emerging European-like porcine reproductive and respiratory syndrome virus isolates in the United States. J Virol 2004; 78: 3684–3703.
11. Murtaugh MP. Update on PRRSV immunology and viral genetics: from hopeless to hopeful, in Proceedings. 40 Annu Meet Am Assoc Swine Vet 2009; 459–462.
12. Holtkamp DJ, Kliebenstein JB, Neumann EJ. Assessment of the economic impact of porcine reproductive and respiratory syndrome virus on United States pork producers. J Swine Health Prod 2013; 21: 72–84.
13. Neumann EJ, Kliebenstein JB, Johnson CD, et al. Assessment of the economic impact of porcine reproductive and respiratory syndrome on swine production in the United States. J Am Vet Med Assoc 2005; 227: 385–392.
14. Christopher-Hennings J, Nelson EA, Nelson JK, et al. Effects of a modified-live virus vaccine against porcine reproductive and respiratory syndrome in boars. Am J Vet Res 1997; 58: 40–45.
15. Christopher-Hennings J, Nelson EA, Nelson JK, et al. Detection of porcine reproductive and respiratory syndrome virus in boar semen by PCR. J Clin Microbiol 1995; 33: 1730–1734.
16. Mengeling WL, Lager KM, Vorwald AC. Clinical effects of porcine reproductive and respiratory syndrome virus on pigs during the early postnatal interval. Am J Vet Res 1998; 59: 52–55.
17. Holtkamp DJ, Torremorell M, Morrison B, et al. Terminology for classifying swine herds by porcine reproductive and respiratory syndrome virus status. J Swine Health Prod 2011; 19: 44–56.
18. Yeske P. A decision tree to guide to answering the question, is this a new virus or not to the farm?, in Proceedings. Allen D. Leman Swine Conf 2013; 71–75.
19. Martin-Valls GE, Kvisgaard LK, Tello M, et al. Analysis of ORF5 and full-length genome sequences of porcine reproductive and respiratory syndrome virus isolates of genotypes 1 and 2 retrieved worldwide provides evidence that recombination is a common phenomenon and may produce mosaic isolates. J Virol 2013; 88: 3170–3181.
20. Muellner P, Zadoks RN, Perez AM, et al. The integration of molecular tools into veterinary and spatial epidemiology. Spat Spatiotemporal Epidemiol 2011; 2: 159–171.
21. Shi M, Lam TT-Y, Hon C-C, et al. Molecular epidemiology of PRRSV: a phylogenetic perspective. Virus Res 2010; 154: 7–17.
22. Shi M, Lam TTY, Hon CC, et al. Phylogeny-based evolutionary, demographical, and geographical dissection of North American type 2 porcine reproductive and respiratory syndrome viruses. J Virol 2010; 84: 8700–8711.
23. Stadejek T, Stankevicius A, Murtaugh MP, et al. Molecular evolution of PRRSV in Europe: current state of play. Vet Microbiol 2013; 165: 21–28.
24. Shi M, Lemey P, Singh Brar M, et al. The spread of type 2 porcine reproductive and respiratory syndrome virus (PRRSV) in North America: a phylogeographic approach. Virology 2013; 447: 146–154.
25. Brar MS, Shi M, Ge L, et al. Porcine reproductive and respiratory syndrome virus in Ontario, Canada 1999 to 2010: genetic diversity and restriction fragment length polymorphisms. J Gen Virol 2011; 92: 1391–1397.
26. Rosendal T, Dewey C, Young B, et al. Distribution of genotypes of porcine reproductive and respiratory syndrome virus in Ontario during 2004–2007 and the association between genotype and clinical signs of disease. Can J Vet Res 2010; 74: 118–123.
27. Wang X, He K, Zhang W, et al. Genetic diversity and phylogenetic analysis of porcine reproductive and respiratory syndrome virus isolates in East China. Infect Genet Evol 2014; 24: 193–201.
28. Xie J, Zhu W, Chen Y, et al. Molecular epidemiology of PRRSV in South China from 2007 to 2011 based on the genetic analysis of ORF5. Microb Pathog 2013; 63: 30–36.
29. Zhang H-Y, Liang J-J, Meng X-M, et al. Molecular epidemiology of PRRSV from China's Guangxi province between 2007 and 2009. Virus Genes 2012; 46: 71–80.
30. Choi H-W, Nam E, Lee YJ, et al. Genomic analysis and pathogenic characteristics of type 2 porcine reproductive and respiratory syndrome virus nsp2 deletion strains isolated in Korea. Vet Microbiol 2014; 170: 232–245.
31. Kvisgaard LK, Hjulsager CK, Brar MS, et al. Genetic dissection of complete genomes of type 2 PRRS viruses isolated in Denmark over a period of 15 years. Vet Microbiol 2013; 167: 334–344.
32. Drigo M, Franzo G, Gigli A, et al. The impact of porcine reproductive and respiratory syndrome virus genetic heterogeneity on molecular assay performances. J Virol Methods 2014; 202: 79–86.
33. Nguyen TT-D, Nguyen TT, Le HT-T, et al. Genetic analysis of ORF5 porcine reproductive and respiratory syndrome virus (PRRSV) isolated in Vietnam. Microbiol Immunol 2013; 57: 518–526.
34. Nilubol D, Tripipat T, Hoonsuwan T, et al. Genetic diversity of the ORF5 gene of porcine reproductive and respiratory syndrome virus (PRRSV) genotypes I and II in Thailand. Arch Virol 2012; 158: 943–953.
35. Kwong GPS, Poljak Z, Deardon R, et al. Bayesian analysis of risk factors for infection with a genotype of porcine reproductive and respiratory syndrome virus in Ontario swine herds using monitoring data. Prev Vet Med 2013; 110: 405–417.
36. Goldberg TL, Weigel RM, Hahn EC, et al. Associations between genetics, farm characteristics and clinical disease in field outbreaks of porcine reproductive and respiratory syndrome virus. Prev Vet Med 2000; 43: 293–302.
37. Badaoui B, Grande R, Calza S, et al. Impact of genetic variation and geographic distribution of porcine reproductive and respiratory